Waveform Processing – Cutting spikes

The first question to answer is why there are these spikes. Should there be spikes in these places, just with a lower amplitude? We cannot tell from the interpolated lines how these spikes look like. Are they single points or are there multiple points above the threshold? Could the be cause by processing issues?

If this data is real, maybe you could just set the y-axis to logarithmic scaling, for example.

As a first step, attach some raw data and explain exactly how you are reading it. We cannot debug these problem from a low resolution picture.

Just coercing spikes to a lower range is irreversible data manipulation that does not address the potential underlying issues.